a tarzan yell for conservation: a new chameleon, calumma ... · madagascar is the ongoing rapid...

167

Calumma tarzan sp. n. from Madagascar

All articles available online at http://www.salamandra-journal.com© 2010 Deutsche Gesellschaft für Herpetologie und Terrarienkunde e.V. (DGHT), Rheinbach, Germany

SALAMANDRA 46(3) 167–179 20 August 2010 ISSN 0036–3375

Introduction

Madagascar is an island that is recognized as one of the richest biodiversity centers with many microendemic taxa and with one of the highest rates of natural habitat loss worldwide (Myers et al. 2000). Current conservation pri-orities in Madagascar are focused on multitaxonomic ap-proaches and no doubts exist that these rather than single-taxon approaches are critical for identifying and protecting areas likely to promote the persistence and conservation of a maximum number of different species (e.g., Kremen et al. 2008). Another concept to increase awareness and pub-licity for an area is to choose a charismatic or impressive species that attracts popular support (“flagship species”, e.g. Williams et al. 2000) to represent a habitat or ecosys-tem in need of conservation, in order to successfully lev-erage the conservation of all species contained therein. In Madagascar, lemurs are typically used as flagship species for conservation activities, besides a few other representa-tives of Madagascar’s enigmatic fauna such as the toma-to frog Dyscophus antongilii in Maroantsetra (Tessa et al. 2007), or some charismatic species of birds (Rabearivony et al. 2010). However, especially the herpetofauna of Mada-

gascar contains more species of striking and colourful ap-pearance (e.g. Glaw et al. 2010, Vieites et al. 2010), with a high recognition value and with great potential to obtain people’s sympathies (e.g. Wollenberg et al. in press).

One of the greatest problems for nature conservation in Madagascar is the ongoing rapid fragmentation of forest habitats and the unstable situation of forests even in some major protected areas due to poor governance and under-mined judicial control (e.g. D’Cruze et al. 2008, Randria-malala & Liu 2010). Recently, illegal large-scale logging of rosewood within protected areas in the wake of the po-litical crisis in 2009 created international awareness and protests (Randriamalala & Liu 2010). In many regions of the island, the forest has been reduced to very small, patchy fragments that are often more or less disturbed. These fragments are no longer suitable for the protection of large mammals (e.g., lemurs of the genera Propithecus or Indri) and, as a consequence, they are often neglected in the delimitation of protected areas. However, a large por-tion of the Malagasy fauna (including several chameleon species) is at least somewhat tolerant to forest disturbance (e.g., Jenkins et al. 1999, 2003, Rabearivony et al. 2007), and it would therefore be feasible to preserve major por-

A Tarzan yell for conservation: a new chameleon, Calumma tarzan sp. n., proposed as a flagship species for the creation of new nature reserves in Madagascar

Philip-Sebastian Gehring1, Maciej Pabijan1,6, Fanomezana M. Ratsoavina1,4,5, Jörn Köhler2, Miguel Vences1 & Frank Glaw3

1) Division of Evolutionary Biology, Zoological Institute, Technical University of Braunschweig, Spielmannstr. 8, 38106 Braunschweig, Germany

2) Department of Natural History – Zoology, Hessisches Landesmuseum Darmstadt, Friedensplatz 1, 64283 Darmstadt, Germany3) Zoologische Staatssammlung München, Münchhausenstr. 21, 81247 München, Germany

4) Departement de Biologie Animale, Universite d’Antananarivo, BP 906. Antananarivo, 101, Madagascar.5) Grewcock´s Center for Conservation Research, Omaha´s Henry Doorly Zoo, 3701 South 10th Street, Omaha,

NE 68107-2200, U.S.A.6) Department of Comparative Anatomy, Institute of Zoology, Jagiellonian University, ul. Ingardena 6, 30-060 Kraków, Poland

Corresponding author: Philip-Sebastian Gehring, e-mail: [email protected]

Manuscript received: 23 June 2010

Abstract. We describe Calumma tarzan sp. n., a morphologically distinct chameleon species of the Calumma furcifer spe-cies group from rainforest fragments in the Anosibe An’Ala region of central eastern Madagascar. Males and females of this species differ from all other species of the Calumma furcifer group by its rostral crests, which are fused anteriorly to form a spade-like ridge that slightly projects beyond the snout tip (less than 1 mm), by a unique stress colouration with a pattern of bright yellow and green, and by significant genetic divergence as assessed by an analysis of sequences of a fragment of the mitochondrial ND4 gene. We suspect that Calumma tarzan might be restricted to a rather small and fragmented distribu-tion range and we consider the IUCN conservation status of “Critically Endangered” as adequate for this species. Moreover, we propose Calumma tarzan as a flagship species for the establishment of new protected areas in central-eastern Madagas-car and suggest that biological surveys be carried out in the largely unexplored forest fragments of the region.

Key words. Squamata, Chamaeleonidae, Calumma tarzan, new species, forest fragments, Madagascar.

168

Philip-Sebastian Gehring et al.

tions of the highly microendemic Malagasy fauna by also protecting small fragments with partly degraded buffer zones.

Madagascar is the center of chameleon species diversi-ty (Brygoo 1971, 1978), with many species being restrict-ed to primary forest habitats. The three Malagasy genera Brookesia, Calumma, and Furcifer include approximately 40% of the world’s known chameleon species, including the largest and the smallest representatives. They are popu-lar and enigmatic components of the Malagasy herpetofau-na for ecotourists (Wollenberg et al. in press), qualifying them as ideal flagship species for conservation activities.

The genus Calumma is currently composed of 30 spe-cies, which are distributed in the humid rainforests and montane regions of eastern and central Madagascar (Glaw & Vences 2007). A further species (C. tigris) is endemic to the continental Seychelles islands, but this species shows little external similarity to any Malagasy species group and its affinities to Calumma are currently under study (T. Townsend et al. unpublished data). In contrast to the ge-nus Furcifer, in which only a few new species have been discovered in recent times (Glaw et al. 2009), nine new species of Calumma have been described in the last 15 years and a further four taxa have been elevated to species rank (Böhme 1997, Andreone et al. 2001, Raxworthy & Nuss-baum 2006).

According to Andreone et al. (2001), the Calumma fur-cifer species group currently consists of the following spe-cies: Calumma furcifer (Vaillant & Grandidier, 1880), C. gastrotaenia (Boulenger, 1888), C. marojezense (Bry-goo, Blanc & Domergue, 1970), C. andringitraense (Bry-goo, Blanc & Domergue, 1972), C. peyrierasi (Brygoo & Domergue, 1974), C. guillaumeti (Brygoo, Blanc & Domergue, 1974), C. glawi Böhme, 1997, C. vatosoa An-dreone, Mattioli, Jesu & Randrianirina, 2001, and C. vencesi Andreone, Mattioli, Jesu & Randrianirina, 2001. With the exception of the two poorly known and more morphologically divergent species C. peyrierasi and C. vatosoa, these species represent a rather homogeneous group with only subtle interspecific differences in exter-nal morphology, but often with rather strong differences in hemipenis morphology (e.g. Böhme 1997, Andreone et al. 2001).

In the following, we describe a new species in the Ca-lumma furcifer species group from the Anosibe An’Ala re-gion that differs by morphology, life colouration and ge-netic differentiation from all other species in the group. Moreover, with the recent discovery of this unique new chameleon species from one of Madagascar’s most defor-ested areas, we want to (1) highlight the potential of pro-tecting small forest remnants in the network of protected areas, (2) draw attention to an area that still remains large-ly unstudied, and (3) create publicity for a species that is probably already on the brink of extinction.

Materials and methods

Most specimens were collected at night during the rainy season, using torches and headlamps to detect roosting chameleons in the vegetation. The type specimens were collected, anesthetized and killed by injection with chlo-

robutanol, fixed with 90% ethanol and stored in 70% etha-nol. Specimens dealt with in this paper were subsequent-ly deposited in the collection of the Zoologische Staatssa-mmlung München (ZSM), Munich, Germany and in the collection of the Universite d’Antananarivo, Departement de Biologie Animale (UADBA), Antananarivo, Madagas-car. Tissue samples of additional specimens were taken by tail clipping freshly collected specimens and stored in 98% ethanol at the Technical University of Braunschweig, Zoo-logical Institute, Braunschweig, Germany. FGMV, FGZC, PSG and MPFC refer to field numbers of F. Glaw/M. Vences, F. Glaw, P.-S. Gehring, and M. Pabijan respec-tively. Locality coordinates were recorded with GPS receiv-ers.

Morphological measurements were taken with digital callipers to the nearest 0.1 mm by P.-S. Gehring. Defini-tion of measurements and the description scheme of the holotype largely follow Andreone et al. (2001). Snout-vent length is abbreviated SVL.

Mitochondrial DNA variation was studied in five speci-mens of the new species, including the holotype, one para-type and three individuals sampled in the forest fragments at Tarzanville and Ambatofotsy. Additionally, we incorpo-rated available tissue samples from other species of the C. furcifer species group. Tissue samples of C. cf. tarzan were collected by FR in a forest fragment close to Marolambo in May 2009. Tissue samples of C. cf. vencesi were collected by MV, FR and David R. Vieites in the Makira forest in north-eastern Madagascar in February 2008. Tissue sam-ples of C. cf. marojezense were collected in a forest frag-ment (Ambodivoahangy) along the Antainambalana river valley in north-eastern Madagascar by PSG, FG, JK, MP and FR in April 2010. Since the taxonomical identifica-tion of these latter specimens is not yet assured, our spe-cific assignment is tentative. Data on three described spe-cies of the C. furcifer group are missing from our analysis: C. guillaumeti, C. peyrierasi and C. vatosoa. After extrac-tion of total genomic DNA, using different standard pro-tocols, fragments of the NADH dehydrogenase subunit 4 gene (ND4) were PCR-amplified using the primers ND4 (5’ CAC CTA TGA CTA CCA AAA GCT CAT GTA GAA GC 3’) and LeutRNA (5’ CAT TAC TTT TAC TTG GAT TTG CAC C 3’), following standard protocols (see Aréva-lo et al. 1994). After purification (EXOSAP), the fragments were resolved on an automated DNA sequencer (ABI 3130 XL Applied Biosystems). Sequences were validated and aligned manually with the software CodonCode Align-er (CodonCode Corporation). Genbank accession num-bers of the newly resolved DNA sequences are HM776645 (FGZC 4510), HM776649 (FGZC 4514), HM776650 (PSG 2114), HM776651 (MPFC 40), and HM776652 (MPFC 97) for Calumma tarzan; HM776646 (PSG 1377), HM776647 (PSG 1376), and HM776648 (PSG 1341) for C. cf. tar-zan from Marolambo; HM776656 (FGMV 2001.590) for C. andringitraense; HM776659 (FGMV 2002.321) for C. glawi; HM776658 (FGMV 2002.987) for C. gastrotaenia; HM776653 (MPFC 110) and HM776657 (FGMV 2001.240) for C. furcifer; HM776660 (ZCMV 11275) for C. cf. venc-esi; HM776654 (PSG 2364) and HM776655 (PSG 2350) for C. marojezense; and HM776661 (FGMV 2002.1221) for C. boettgeri. A neighbor joining tree was constructed in Mega 4.0 (Tamura et al. 2007) based on distances calcu-

169


lated with the Tamura-Kumar modification of the LogDet method (Tamura & Kumar 2002) due to the composition-al heterogeneity between Calumma gastrotaenia, C. glawi and C. tarzan sp. n. Support values for nodes were gen-erated by bootstrapping (10000 replicates). A maximum parsimony analysis with unordered and equally weighted characters was carried out in PAUP 4.0b10 (Swofford 2003). A heuristic search was conducted with 100 random taxon stepwise addition sequences and tree bisection re-connection (TBR) branchswapping. The topology was re-constructed using the 50% majority rule consensus with support values assessed by 10000 bootstrap pseudorepli-cates. We also applied a model-based analysis in a Bayesian framework. Model selection was carried out in MrModel-Test v2.3 (Posada & Crandall 1998, Nylander 2004). The Akaike information criterion used in this software se-lected GTR+I+G as the best-fit model of nucleotide sub-stitution. We implemented this model in MrBayes v3.1.2, and programmed two runs with four chains that were run for a total of 5 million generations and sampled every 100 generations. The two runs converged into a stationary dis-tribution after 10000 generations, for which reason we dis-carded the first 25% of the generations as burn-in, leaving 37500 generations from which parameter values were sum-marized and a majority rule consensus tree was produced with posterior probabilities calculated as the frequency of samples recovering each clade (Huelsenbeck & Ron-quist 2001).

Systematics

Calumma tarzan sp. n.Figs. 1–5

Holotype: ZSM 219/2010 (FGZC 4510), adult male with in-completely everted hemipenes, collected in the forest frag-ment near Tarzanville (19°19’50.3’’ S, 48°13’22.5’’ E, 847 m above sea level), Moramanga Province, central eastern Madagascar, on 12 April 2010 by P.-S. Gehring, F. Glaw, J. Köhler, K. Mebert, E. Rajeriarison & F. M. Ratsoavi-na.

Paratypes: UADBA uncataloged (FGZC 4511), UADBA un-cataloged (FGZC 4513), ZSM 221/2010 (FGZC 4515), three adult males; UADBA uncataloged (FGZC 4405), UADBA uncataloged (FGZC 4512), ZSM 220/2010 (FGZC 4499), and ZSM 222/2010 (FGZC 4514), four adult females; UAD-BA uncataloged (FGZC 4519), one juvenile specimen, all with same collection data as the holotype.

Additional material: Additional tissue samples (without corresponding voucher specimens) for genetic analyses were available from a population of C. tarzan from the type locality Tarzanville (19°19’50.3’’ S, 48°13’22.5’’ E, 847 m a.s.l.), PSG 2114, 2116, 2117, 2130, 2133, 2138, 2139, 2142, 2151; from a forest fragment less than one kilometer north of Tarzanville (19°19.461’ S, 48°13.193’ E, 881 m a.s.l.), PSG 2178–2181; and from Ambatofotsy approximately 25 km southeast of Tarzanville (19°32’35.2” S, 48°18’59.5” E, 907 m a.s.l.), MPFC 38–42, 45, 94. In 2009, FR collected tissue samples from several juveniles of an unidentified chamele-

on species of the C. furcifer group in a forest fragment close to Marolambo (20°03’27.3” S, 48°11’00.0” E, 670 m a.s.l.), which in our analyses turned out to represent the sister group to C. tarzan and we therefore assign this population to C. cf. tarzan (PSG 1341–1343, 1351, 1368, 1376–1378, 1386).

Diagnosis: A medium-sized green chameleon (snout-vent length 61–72 mm, total length 119–150 mm) with a charac-teristic rostral profile, consisting of the rostral crests which are fused anteriorly to form a spade-like ridge that slightly projects beyond the snout tip (less than 1 mm), and a dis-tinct colour pattern in life, consisting of a green to bright yellow ground colouration (with dark transversal bands when stressed) and a brown to blackish casque and neck in males. The following combination of characters leads us to assign the new species to the C. furcifer group: the small to moderate adult body size, the absence of occipital lobes (although small rudimentary lobes are present in males of C. glawi), a very low casque, absence of gular and ven-tral crests, and a more or less uniformly green colouration usually with a dark line running through the eyelids, and a broad brownish mid-ventral band that is bordered by a narrower white band on each side (except for C. peyrierasi that has only a single whitish ventral band). Calumma tar-zan differs from all the other species in this group by the unique “spade-like” rostral profile (the two rostral crests unite anteriorly on the snout forming a rectangular ridge when viewed from above instead of a pointed snout tip in most other species of the group), and the blackish-brown casque and neck in males. Beside these unique character-istics, males of C. tarzan differ from those of C. furcifer by the absence of a dorsal crest and the absence of a bifid ros-tral appendage; from C. gastrotaenia by the much lower casque, absence of a dorsal crest and absence of a white lat-eral band; from C. guillaumeti by the lower casque, the ab-sence of a dorsal crest and absence of a white lateral band; from C. andringitraense by a lower casque and the absence of a white lateral band with two rows of white spots along the flanks; from C. glawi by the absence of occipital lobes and absence of a yellow lateral band; from C. marojezense by the absence of a white lateral band and presence of only one light spot on the flank (instead of two); from C. peyri-erasi by a much lower casque, the absence of a dorsal crest and absence of reddish inner sides of legs; from C. vatosoa by the absence of a parietal crest, absence of a vein-like dor-sal pattern and absence of a large orange spot at mid-flank; from C. vencesi by the absence of a nuchal fold and absence of a dorsal crest. The females of C. tarzan differ from the females of the other species (as far as they are known and with the possible exception of C. furcifer) by the rectangu-lar rostral profile which, however, is less pronounced than in the males. In addition, C. tarzan differs from C. andrin-gitraense, C. furcifer, C. gastrotaenia, C. glawi and C. maro-jezense by substantial genetic differentiation (Fig. 7).

Description of the holotype: Adult male, in good condi-tion, both hemipenes incompletely everted; SVL 72.0 mm; tail length 77.8 mm; further morphological measurements are provided in Table 1 (characters according to Tables 1 and 2 in Andreone et al. 2001). Head without rostral ap-pendage, but with a unique “spade-like” rostral profile, the two rostral crests uniting anteriorly into a thickened

170


Figure 1. Male holotype of Calumma tarzan sp. n. (ZSM 219/2010) in life. (A) Lateral view when relaxed, (B) moderate stress coloura-tion, and (C) stress colouration fully developed.

ridge, forming together a rectangular ridge in dorsal view that slightly (clearly less than 1 mm) projects beyond the snout tip; rostral crests almost parallel, but slightly diverg-

ing posteriorly, distance between anterior tips 3.5 mm; pa-rietal crest entirely absent; supra-orbital crest rounded in lateral view and formed by a single, rather smooth row of

171


Figure 2. Head and neck of a male Calumma tarzan sp. n. in dorsal view, showing the characteristic brownish colour pattern and the spade-like ridge projecting beyond the snout tip.

tubercles; lateral crest moderately distinct, smooth in lat-eral view; temporal crest indistinct, both crests fusing pos-teriorly; parietal crest absent; no traces of occipital lobes; no traces of gular and ventral crests; dorsal crest absent. Body laterally compressed with fine homogeneous scala-tion, except in the vicinity of the cranial crests and neck region where scales are slightly larger; axillary pits distinct, deeply recessed; limbs and tail with homogeneous scala-tion, tail without dorsal crest, feet without tarsal spines; tail base moderately swollen.

Hemipenes small, both most likely incompletely everted (total length on right side 4.8 mm, on left side 5.5 mm). The general morphology (Fig. 6) differs from that of other spe-cies in the Calumma furcifer group (see drawings and de-scriptions in Böhme 1997 and Andreone et al. 2001) and is therefore difficult to interpret and perhaps partially an artifact. The truncus, especially in asulcal view, is covered with numerous distinct papillae (unknown from the oth-er species) whereas calyces are not recognizable (present in the other species). Apex with one distinct and one in-distinct pair of rotulae, without clearly recognizable struc-tures in the area between the four rotulae, which are typical for the other species of the C. furcifer group (see Böhme 1997).

In life, when relaxed, the male holotype (Fig. 1) exhib-ited a rather light green colouration with a yellow ground colour during the day. Starting at the snout, the casque and roughly the anterior third of the dorsum were covered with black to charcoal-colour delimited by a sharp line along the casque and the lateral crest to the rostral crest and on to the upper tip of the snout (Fig. 2). Along the back in posteri-or direction the blackish colour faded to whitish-grey. One single, small, light yellowish spot (1.4 mm in diameter) on the anterior flanks. No visible light lateral band, which is in contrast to most other species of the C. furcifer group. A dark horizontal line through the eyelid, extending hori-zontally in a posterior direction to the insertion of the fore-limbs. Belly whitish-green, with a broad, grey-brown lon-gitudinal band from the throat to the cloaca broadest mid-ventrally (3.4 mm) and narrower towards the forelimbs, where it fades out in irregular turquoise and whitish spots. This broad band is fringed by an indistinct reddish line that is bordered on either side with a white band ca. 1 mm wide. These bands merge with the white ground colour of the throat at level of the insertion of forelimbs. The pattern of a central brown band bordered with a narrow white band on either side continues onto the tail, but fades within the first half of the tail. The inner mucous integument of the mouth is darkly pigmented; axillary pockets are whitish. Iris cop-per-coloured. After being handled, the holotype showed a unique stress colouration (Fig. 1) unknown from other species of the C. furcifer group. The yellow ground colour became prominent and interrupted the green coloura-tion by oblique transversal bands on body and tail. Sev-eral irregular yellow spots appeared on head, flanks, and extremities. The eyelids became almost completely yellow and the tail almost completely yellow with transversal dark green bands. There was no change in the dark colouration of the head, but the greyish patch on the shoulders became more whitish. In preservative, the holotype faded and the green colouration almost disappeared, changing into greyish-blue. Dark colouration of the head indistinguish-

able from the dark dorsal colouration of the flanks and the back. Dorsum dark bluish, turning into lighter shades ven-trally. One single light spot on each flank, axillary pockets white. Ventrally, the medioventral band has faded to grey, still bordered by two narrow white bands, composed of ir-regular dark spots on the throat extending until the ante-rior half of the tail.

Variation: Description of the female paratype ZSM 222/2010 (FGZC 4514): Adult female, in good condition, but with a longitudinally cut venter; SVL 61.4 mm; tail 57.9 mm; further morphological measurements are provided in Table 1. Head without rostral appendage, but with a char-acteristic “spade-like” rostral profile, less pronounced than in the male; the two rostral crests unite anteriorly on the snout into a thickened ridge, forming together a rectangu-lar ridge in dorsal view that very slightly projects beyond the snout tip; rostral crests almost parallel, but slightly di-vergent posteriorly, distance between anterior tips 1.9 mm; parietal crest entirely absent; supra-orbital crest rather in-distinct; lateral crest distinct, smooth in lateral view; tem-poral crest indistinct, both crests fuse posteriorly; parietal crest absent; no traces of occipital lobes; no traces of gular and ventral crests; dorsal crest absent. Body laterally com-pressed with fine homogeneous scalation except on the

172


Figure 3. Female paratype of Calumma tarzan sp. n. (ZSM 222/2010) in life. (A) Lateral view when relaxed, and (B, C) differently developed stress colourations.

173


head, which is covered with slightly larger scales; axillary pits distinct; limbs and tail with homogeneous scalation, tail without dorsal crest, feet without tarsal spines; tail base not swollen.

For measurements of ZSM type specimens see Table 1. Specimens deposited in UADBA were not available for de-tailed examination. Females of C. tarzan have a distinctly less swollen tail base.

Concerning colouration in life, males in a relaxed state were rather uniform in their colour patterns. The brown colour on the head and neck can extend posteriorly on the dorsal ridge to midbody. Depending on agitation, several irregular dark spots on the head, flanks, and the extremi-ties can be visible, as well as dark oblique transversal bands and a more pronounced dark horizontal line through the eyelid which can extend further posteriorly than in a re-laxed state. The yellow spot on the flank can be surrounded by a dark line.

Life colouration of the females is uniformly light green (in a relaxed state) to bright yellow with a network of darker horizontal stripes and irregular darker spots on the flanks in a stressed state (Fig. 3, photographed during the day). A dark horizontal line through the eyelid is also present in females.

Juveniles showed more or less the same colouration as the females; uniformly light green to yellow, although the light spot on the flanks was clearly recognizable in juvenile males (Fig. 4, photographed at night).

Etymology: The species was discovered in a small rainfor-est fragment very close to the village formerly known as Tarzanville (since recently called Ambodimeloka, meaning “place of the curve”). Therefore, we dedicate the new spe-cies to the fictional forest man “Tarzan” in the hope that this famous name will promote awareness and conserva-tion activities for this apparently highly threatened new species and its habitats, in the mid-altitude rainforest. We consider this epithet as an invariable noun in apposition.

Distribution, habitat and habits: Calumma tarzan is known only from its type locality (a forest fragment of about 1.5 km in length and less than 0.3 km in width), another small and degraded rainforest fragment just less than one kilom-eter north of Tarzanville (19°19.461’ S, 48°13.193’ E, 881 m a.s.l.) and from Ambatofotsy (19°32’35.2” S, 48°18’59.5” E, 907 m a.s.l.), a forest fragment (of ca. 6 km in length and 1.5 km in width) proposed as a potential protected area, lo-cated approximately 25 km southeast of Tarzanville (Fig. 8).

The forests surrounding these three known localities com-prise a total area of less than 10 km2 and are situated along the slopes of the Manambolo river valley, which runs from north to south before discharging into the Mangoro riv-er. All voucher specimens were collected inside the forest fragment close to Tarzanville.

A survey in 2009 in a small forest fragment in the Maro-lambo area some 50 km south of Ambatofotsy and situat-ed south of the Mangoro river (Fig. 8) yielded juveniles of an unidentified form of the C. furcifer group, showing per-suasive similarities with C. tarzan (Fig. 4b) which there-fore may also be assignable to the new species. Only tissue samples and no voucher specimen were collected from this population. We here tentatively refer to this Marolambo population as C. cf. tarzan. Further studies and adult spec-imens are necessary to assess the taxonomic status of this population.

All known localities of C. tarzan are within an altitudi-nal range of ca. 800–910 m above sea level. All individuals were found in mid-April. At that time they were relatively abundant and found roosting in the vegetation at night, ca. 1–4 m above the forest floor. Small and probably recently

Figure 4. Juveniles of (A) Calumma tarzan sp. n. from the type locality, and (B) Calumma cf. tarzan from the Marolambo region.

Table 1. Biometric measurements (in mm) of available type specimens of Calumma tarzan sp. n. from the forest fragment close to Tarzanville. Definition and selection of characters according to Andreone et al. (2001); abbreviations as follows: SVL, snout-vent length; TL, tail length; HL, head length; HD, head depth; HW, head width; SD, socket diameter (measured horizontally); AGD, axilla-groin distance; HT, holotype; PT, paratype.

Voucher number Status Sex SVL TL HL HD HW SD AGD

ZSM 219/2010 HT male 72.0 77.8 19.4 12.6 9.6 5.1 38.0

ZSM 220/2010 PT female 60.6 58.3 20.0 12.4 8.8 4.1 32.3ZSM 221/2010 PT male 68.1 72.4 21.6 12.1 9.2 5.6 39.4ZSM 222/2010 PT female 61.4 57.9 19.9 12.2 8.7 4.6 34.3

174


hatched juveniles were much more common than adults and seemingly perching at lower positions (mostly ca. 1–2 above the forest floor). One female (ZSM 222/2010) con-tained three large eggs (diameters 11.0 x 6.3, 10.7 x 6.2, and 10.7 x 6.3 mm). Another female (FGZC 4405) was found at night hanging head down dead in a tree, attached with the tip of the tail to a twig (Fig. 5). Although it apparently had died at most a few hours before (its colour was still similar to that in life), many large unidentified insect larvae were found in its body the next morning, suggesting that the lar-vae had already infected the living chameleon and might have caused its death.

Molecular differentiation and phylogenetic relationships: The alignment of partial DNA sequences of the ND4 gene consisted of 17 sequences with 537 aligned nucleotide po-sitions. A total of 191 nucleotides were variable, 134 of which were parsimony informative. The heuristic search in the MP analysis produced a single most parsimonious tree with a length of 376 (consistency index, CI=0.628; re-tention index, RI=0.702). The BI runs resulted in identi-cal topologies and very similar likelihood estimates (mean lnL= -2193.10).

The ND4 gene sequences of C. tarzan showed a high divergence to all included taxa of the C. furcifer species group (Fig. 7). Six clades were highly supported by dis-tance, discrete character, and model-based phylogenetic analyses (Fig. 7). Specimens assigned to C. tarzan from the type locality and Ambatofotsy formed a highly supported clade. No sequence variation was detected within C. tar-zan. The sister group to C. tarzan was C. cf. tarzan from Marolambo, differing by an average of 18.3 nucleotide sub-stitutions (3.5% uncorrected p-distance) and 2 aminoacid substitutions. The average divergence of C. tarzan from the other nominal species of the C. furcifer group includ-ed in our analysis was 58.5–73.0 substitutions (12.6–15.7% p-distance), but phylogenetic relationships between most of the other taxa incorporated into the analysis are unre-solved. However, the main purpose of this tree is not to re-liably establish relationships within this group, but to give an indication of the molecular differentiation between taxa.

Available names: Two junior synonyms are currently rec-ognized in the Calumma furcifer species group and need to be discussed as possible earlier available names for C. tarzan: Chamaeleon grandidieri Mocquard, 1900 (syno-nymized with C. gastrotaenia by Werner 1902), was based on a single female holotype (MNHN 99416) from the sur-roundings of Suberbieville (today named Maevatanana ac-cording to Viette 1991) in central north-western Mada-gascar, which is an unusual locality for a species of the ge-nus Calumma as these are mostly restricted to humid east-ern Madagascar. The species was described in two differ-ent papers. One of the descriptions (Mocquard 1900a) is very short, comprising just seven lines, whereas the second (Mocquard 1900b) is much more detailed, and includes an illustration. According to these descriptions, the holo-type had a SVL of 45 mm, a tail length of 43 mm, an axil-lary mite pocket, and a homogeneous pholidosis, but no casque, no dorsal, gular and ventral crests, and no crests on the head. An evaluation of the validity of this taxon is dif-ficult since the original description of C. gastrotaenia was based on a male and a subadult (Boulenger 1888) and with the imprecise type locality “Madagascar”. The status of C. grandidieri was not discussed or reassessed by recent authors who described new species in the C. furcifer group (Böhme 1997, Andreone et al. 2001). However, since the type locality of C. grandidieri is in a different biogeograph-ic and climatic region than that of C. tarzan, it is highly un-likely that both taxa are conspecific.

Chamaeleon chauvini was described by Methuen & Hewitt (1913) based on a male holotype collected near Tamatave (today Toamasina). The authors considered this taxon as closely related to C. furcifer and mentioned the possibility that “it may eventually prove to be a juvenile of that species”. From the description and the accompanying photograph of the holotype, a dorsal crest of small tuber-cles and a bony rostral process projecting about 1 mm be-yond the upper lip and being bifid at the end, are evident. These two characters differ substantially from Calumma tarzan and agree entirely with C. furcifer, leaving no doubts that the synonymy of C. chauvini with C. furcifer is fully justified.

Figure 5. Adult female of Calumma tarzan sp. n. (UADBA/FGZC 4405), found hanging dead in a tree at the type locality.

175


DiscussionRelationships

With Calumma tarzan, we add a new morphologically dis-tinct chameleon species to the C. furcifer species group. Most species currently allocated to this group form a rela-tively homogeneous morphological group, whereas C. pey-rierasi and C. vatosoa are morphologically more divergent, indicating that the phylogenetic relationships within this group and with other related groups (e.g., the Calumma nasutum group) are still poorly understood. Our phylo-genetic analysis of the species group (Fig. 7) based on the ND4 gene fragment did not properly resolve relationships among the species studied either. Nevertheless, it became at least clear that the closest relative of C. tarzan is a barely studied population of morphologically similar chameleons from Marolambo which is distinctly differentiated geneti-cally, however. Further studies may reveal that this Maro-lambo population either represents a separate species, or a differentiated but conspecific population of C. tarzan.

On a larger scale, recent molecular studies suggest that the genus Calumma is polyphyletic (e.g., Raxworthy et al. 2002, Townsend & Larson 2002). The relationships be-tween the species groups are not yet sufficiently resolved and even the monophyly of several species groups in their current definition is still questionable. It is thus likely that a future partition of the genus Calumma will be necessary to properly reflect the phylogenetic history of these chame-leons. In this case, only the clade including the type species of Calumma, C. cucullatum, is likely to retain its current generic name in the future. This type species was found to be the sister species of the C. furcifer group (C. gastro-taenia, C. guillaumeti, andringitraense and marojezense) by Raxworthy et al. (2002), suggesting that even a revised genus Calumma would include the species of the C. fur-cifer group.

Conservation

Several areas around the known distribution range of C. tarzan were surveyed but did not reveal additional locali-ties for the new species: in forests around Niagarakely, situ-ated north of Tarzanville and at a higher altitude (19°13.127’ S, 48°13.930’ E, 1034 m a.s.l.; Fig. 8), we found only a sin-gle specimen of Calumma cf. gastrotaenia. The forest rem-nants between the Manambolo and Mangoro river valleys are mostly above 900 m altitude and therefore are probably not populated by the new species. The same is true for the forests west of the Mangoro River, including an area east of Tsinjoarivo surveyed by us in 2010, in which C. tarzan was not recorded. Surveys in forest fragments of littoral forests in an area between Vatomandry, Mahanoro, and the estu-ary of the Mangoro River did not yield any C. tarzan or any other member of the C. furcifer group (Gehring et al. submitted).

Perhaps the best chances for discovering additional pop-ulations of C. tarzan may be in areas to the east and north-east of the type locality where small fragments of mid-al-titude rainforests still exist (Fig. 8). The available data in-dicate that the distribution of C. tarzan is rather patchy.

Although the real limits of its distribution still need to be assessed, this distinctive species was apparently not discov-ered earlier by other herpetologists or ecotourists, suggest-ing a relatively small range. Furthermore, its tolerance of habitat disturbance might be limited, and a high level of recent logging and degradation was evident in and around the known habitats.

Using these data to assess the IUCN Conservation sta-tus, C. tarzan should be considered as “Critically Endan-

Figure 6. Left hemipenis of the holotype of Calumma tarzan sp. n., (A) asulcal view, (B) sulcal view. Note that the hemipenis is most likely not fully everted.

176


Figure 7. Bayesian phylogram of the Calumma furcifer species group, based on the mitochondrial ND4 gene (GTR+I+G model of nucleotide evolution). Values above nodes indicate Bayesian clade credibility (5 mio. generations), bootstrap support for the neighbor-joining analysis (10000 replicates), and bootstrap support for the maximum parsimony analysis (10000 replicates). Only values above 75% bootstrap support and 0.7 posterior probability are shown.

gered” because its Area of Occupancy (AOO) is probably less than 10 km2, all individuals are in only two major lo-cations, there is continuing decline in the extent and qual-ity of much of its habitat, and none of the habitats are in-

cluded in the network of protected areas yet. In the large area between Analamazoatra-Zahamena-Ankeniheny in the north and the Ranomafana National Park in the south, extending for about 250 km, no low- and mid-altitude ar-

177


eas are protected or proposed for protection (Kremen et al. 2008), resulting in a major “reserve gap” in central eastern Madagascar. One reason for this situation might be that this area is not comprised of any large block of primary for-est and therefore might appear less important for the con-servation of larger animals such as diurnal lemurs. How-ever, at least from a herpetological point of view, even tiny forest fragments can play a crucial role in the conservation of vital populations of different endangered amphibian and reptile species (see e.g. Vallan 2000) and this most likely also holds true for major parts of the invertebrate fauna of such fragments. Preserving small fragments seems of par-ticular importance in areas where intensive slash-and-burn agriculture has destroyed most of the primary habitats, and where at the same time microendemic ranges of certain species were already identified (compare e.g. D’Cruze et al. 2007, Wollenberg et al. 2008). The forest fragments in the Anosibe An’Ala region may perfectly qualify as a con-servation priority. This region has been (surprisingly) very poorly surveyed until now and recent discoveries suggest a

significant degree of microendemism in this area, exempli-fied by the recently described Phelsuma hoeschi (Berghof & Trautmann 2009) or the presence of Calumma sp. aff. gallus, currently under description (P.-S. Gehring et al. unpublished data). The discovery of Calumma tarzan sup-ports the assumption that this area may hold even more microendemic species. Two major conclusions should be drawn from this situation:

(1) The known habitats of this new chameleon, espe-cially Ambatofotsy and the forest fragment at Tarzanville, should be placed under formal protection (either by the Malagasy government or NGOs) as soon as possible in or-der to ensure the survival of this splendid species and other locally endemic species expected to occur in these forest remnants, using C. tarzan as a flagship species. The forest near Tarzanville is very small and very close to the road connecting Moramanga and Anosibe An’Ala. Due to the small size of the fragment, its protection would not require exhaustive financial resources and ensure the survival of a vital population of C. tarzan and many other forest-dwell-

Fig. 8. Modified map of central eastern Madagascar taken from the Madagascar Vegetation Mapping Project (www.vegmad.org), showing the potential distribution range of Calumma tarzan and Calumma cf. tarzan in central eastern Madagascar. The known localities are indicated by arrows, 1) forest close to Tarzanville, 2) Ambatofotsy and 3) forest fragment close to Marolambo; potential localities are surrounded by red circles (a – d). Green colour represents remaining forest areas; bright yellow secondary vegetation; blue colour water bodies.

178


ing species. In addition, this forest could be interesting enough for the establishment of an ecotourist destination, due to its relatively easy accessibility from one of Madagas-car’s most famous and most visited nature reserves (only 2.5 hours drive from Moramanga and 3 hours from An-dasibe). Even if only a limited number of tourists would visit this forest, the income generated by this ecotourism would probably be higher than that obtained by clearing the forest and planting rice or other crops on the slopes, as can be observed in the close vicinity of the habitat of C. tarzan. Conservation activities for this forest are urgent, as it is already exploited for precious timber, and recently established properties are already marking its edges. The second site that C. tarzan was recorded from, Ambatofot-sy, encompasses approximately 9 km2 on a forested plateau between 800–1060 m a.s.l. This forest patch is at present accessible only by foot and therefore more remote than the Tarzanville forest. Moreover, access is hampered by steep rock cliffs partly surrounding the forest, which have to some degree protected the site from logging. Due to its larger size, Ambatofotsy may be a stronghold for the spe-cies, especially in the event of continued habitat degrada-tion at the type locality.

(2) Since we suspect that the level of microendemism in the above mentioned “reserve gap” may be strongly un-derestimated, biological inventory surveys should be car-ried out as soon as possible to improve our understanding of the biodiversity and biogeography of this region. These surveys should also include the small mid-altitude forest fragments east and southeast of Tarzanville, and the Maro-lambo region (see Fig. 8), as these areas have the highest potential of harbouring additional populations of C. tarzan or undescribed related taxa.

Acknowledgements

We are grateful to numerous colleagues, friends and organiza-tions: Jason Lee Brown, Konrad Mebert, Emile Rajeriari-son, Solohery Rasamison, and Francois Randrianasolo helped in the field, and Ruth Kühbandner prepared the hemi-penis drawings. Richard Jenkins (Madagasikara Voakajy) drew our attention to the forest of Ambatofotsy, and the local NGO Mbarakaly (Anosibe An’Ala) helped with logistics. The work of PSG was supported by a project of the evolutionary biology fund-ing initiative of the Volkswagen Foundation. FR was supported by a project of the Volkswagen Foundation in the program “Knowl-edge for Tomorrow – Cooperative Research Projects in Sub-Sa-haran Africa”. MP is supported by a postdoctoral fellowship of the Alexander von Humboldt Foundation. Financial support was also provided by the “Wilhelm-Peters-Fonds” of the Deutsche Gesellschaft für Herpetologie und Terrarienkunde (DGHT). Re-search was carried out in the framework of established collabora-tion with the UADBA. We are grateful to the Malagasy authorities for providing research and export permits as well as to the Ger-man authorities for issuing import permits.

References

Andreone, F., F. Mattioli, R. Jesu & J. E. Randrianirina (2001): Two new chameleons of the genus Calumma from north-east Madagascar, with observations on hemipenial morphology in the Calumma furcifer group (Reptilia, Squa-mata, Chamaeleonidae). – Herpetological Journal, 11: 53–68.

Arévalo, E., S. K. Davis & J. W. Sites, Jr. (1994): Mitochondri-al DNA sequence divergence and phylogenetic relationships among eight chromosome races of the Sceloporus grammicus complex (Phrynosomatidae) in Central Mexico. – Systematic Biology, 43: 387–418.

Berghof, H.-P. & G. Trautmann (2009): Eine neue Art der Gattung Phelsuma Gray, 1825 (Sauria: Gekkonidae) von der Ostküste Madagaskars. – Sauria, 31: 5–14.

Böhme, W. (1997): Eine neue Chamäleonart aus der Calumma gastrotaenia-Verwandtschaft Ost-Madagaskars. – Herpetofau-na, 19(107): 5–10.

Boulenger, G. A. (1888): Descriptions of new reptiles and bat-rachians from Madagascar. – Annals and Magazine of Natural History, ser. 6, 1: 101–107.

Brygoo, E. R. (1971): Reptiles Sauriens Chamaeleonidae – Genre Chamaeleo. – Faune de Madagascar, 33: 1–318.

Brygoo, E. R. (1978): Reptiles Sauriens Chamaeleonidae – Genre Brookesia et complement pour le genre Chamaeleo. – Faune de Madagascar, 47: 1–173.

D’Cruze, N., J. Köhler, M. Franzen & F. Glaw (2008): A con-servation assessment of the amphibians and reptiles of the Forêt d’Ambre Special Reserve, north Madagascar. – Mada-gascar Conservation & Development, 3: 44–54.

D’Cruze, N., J. Sabel, K. Green, J. Dawson, C. Gardner, J. Robinson, G. Starkie, M. Vences & F. Glaw (2007): The first comprehensive survey of amphibians and reptiles at Mon-tagne des Français, Madagascar. – Herpetological Conserva-tion and Biology, 2: 87–99.

Gehring, P.-S., F. M. Ratsoavina & M. Vences (submitted): Filling the gaps – Amphibian and reptile records from low-land rainforests in eastern Madagascar. – Salamandra.

Glaw, F., P.-S. Gehring, J. Köhler, M. Franzen & M. Ven ces (2010): A new dwarf species of day gecko, genus Phelsuma, from the Ankarana pinnacle karst in northern Madagascar. – Salamandra, 46: 83–92.

Glaw, F., J. Köhler & M. Vences (2009): A distinctive new spe-cies of chameleon of the genus Furcifer (Squamata: Chamaele-onidae) from the Montagne d’Ambre rainforest of northern Madagascar. – Zootaxa, 2269: 32–42.

Glaw, F. & M. Vences (2007): A field guide to the amphibians and reptiles of Madagascar. Third edition. Vences & Glaw Ver-lag, 496 pp.

Huelsenbeck, J. P. & F. Ronquist (2001): MRBAYES: Bayesian inference of phylogeny. – Bioinformatics, 17: 754–755.

Jenkins, R. K. B., L. D. Brady, M. Bisoa, J. Rabearivony & R. Griffiths (2003): Forest disturbance and river proximity in-fluence chameleon abundance in Madagascar. – Biological Conservation, 109: 407–415.

Jenkins, R. K. B., L. D. Brady, K. Huston, J. L. D. Kauffmann, J. Rabearivony, G. Raveloson & J. M. Rowcliffe (1999): The population status of chameleons within Ranomafana National Park, Madagascar, and recommendations for future monitor-ing. – Oryx, 33: 38–46.

Kremen, C., A. Cameron, A. Moilanen, S. Phillips, C. D. Tho-mas, H. Beentje, J. Dransfield, B. L. Fisher, F. Glaw, T. C. Good, G. J. Harper, R. J. Hijmans, D. C. Lees, E. Louis Jr., R. A. Nussbaum, C. J. Raxworthy, A. Razafimpahanana, G. E. Schatz, M. Vences, D. R. Vieites, P. C. Wright & M. L. Zjhra (2008): Aligning conservation priorities across taxa in Madagascar, with high-resolution planning tools. – Science, 320: 222–226.

Methuen, P. A. & J. Hewitt (1913): On a collection of reptiles from Madagascar made during the year 1911. – Annals Med-edelingen Transvaal Museum, 3: 183–193.

179


Mocquard, F. (1900a): Diagnoses d’espèces nouvelles de reptiles de Madagascar. – Bulletin du Museum national d’Histoire na-turelle, Paris, 6(7): 345–348.

Mocquard, F. (1900b): Nouvelle contribution à la faune herpe-tologique de Madagascar. – Bulletin de la Societe Philoma-tique de Paris, (9)2: 93–111.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca & J. Kent (2000): Biodiversity hotspots for conser-vation priorities. – Nature, 403: 853–858.

Nylander, J. A. A. (2004): MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala Univer-sity.

Posada D. & K. A. Crandall (1998): Modeltest: testing the model of DNA substitution. – Bioinformatics, 14: 817–818.

Rabearivony, J., L. D. Brady, R. K. B. Jenkins & O. R. Ravoa-hangimalala (2007): Habitat use and abundance of a low-altitude chameleon assemblage in eastern Madagascar. – Her-petological Journal, 17: 247–254.

Rabearivony, J., R. Thorstrom, L.-A. R. de Roland, M. Ra-kotondratsima, T. R. A. Andriamalala, T. S. Sam, G. Ra-zafimanjato, D. Rakotondravony, A. P. Raselimanana & M. Rakotoson (2010): Protected area surface extension in Madagascar: do endemism and threatened species remain useful criteria for site selection? – Madagascar Conservation & Development, 5: 35–47.

Randriamalala, H. & Z. Liu (2010): Rosewood of Madagascar: between democracy and conservation. – Madagascar Conser-vation & Development, 5: 11–22.

Raxworthy, C. J., M. R. J. Forstner & R. A. Nussbaum (2002): Chameleon radiation by oceanic dispersal. – Nature, 415: 784–487.

Raxworthy, C. J. & R. A. Nussbaum (2006): Six new species of occipital-lobed Calumma chameleons (Squamata: Chamaele-onidae) from montane regions of Madagascar, with a new description and revision of Calumma brevicorne. – Copeia, 2006: 711–734.

Swofford, D. L. (2003): PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. – Sinauer Asso-ciates, Sunderland, Massachusetts.

Tamura, K. & S. Kumar (2002): Evolutionary distance estima-tion under heterogeneous substitution pattern among lineag-es. – Molecular Biology and Evolution, 19: 1727–1736.

Tamura, K., J. Dudley, M. Nei & S. Kumar (2007): MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. – Molecular Biology and Evolution, 24: 1596–1599.

Tessa, G., F. M. Guarino, C. Giacoma, F. Mattioli & F. An-drone (2007): Longevity and body size in three populations of Dyscophus antongilii (Microhylidae, Dyscophinae), the to-mato frog from north-eastern Madagascar. – Acta Herpeto-logica, 2: 139–146.

Townsend, T. & A. Larson (2002): Molecular phylogeny and mi-tochondrial genomic evolution in the Chamaeleonidae (Rep-tilia, Squamata). – Molecular Phylogenetics and Evolution, 23: 22–36.

Vallan, D. (2000): Influence of forest fragmentation on amphib-ian diversity in the nature reserve of Ambohitantely, highland Madagascar. – Biological Conservation, 96: 31–43.

Vieites, D. R., F. M. Ratsoavina, R. D. Randrianiaina, Z. T. Nagy, F. Glaw & M. Vences (2010): A rhapsody of colours from Madagascar: discovery of a remarkable new snake of the genus Liophidium and its phylogenetic relationships. – Sala-mandra, 46: 1–10.

Viette, P. (1991): Principales localites où des Insectes ont ete re-cueillis à Madagascar. – Faune de Madagascar, Supplement 2, 88 pp.

Werner, F. (1902): Prodromus einer Monographie der Chamäle-onten. – Zoologische Jahrbücher, 15: 295–460.

Williams, P. H., N. D. Burgess & C. Rahbek (2000): Flagship species, ecological complementarity and conserving the diver-sity of mammals and birds in sub-Saharan Africa. – Animal Conservation, 3: 249–260.

Wollenberg, K. C., R. K. B. Jenkins, R. Randrianavelona, R. Rampilamanana, M. Ralisata, A. Ramanandraibe, O. R. Ravoahangimalala & M. Vences (in press): On the shoul-ders of lemurs: pinpointing the ecotouristic potential of Mada-gascar’s unique herpetofauna. – Journal of Ecotourism.

Wollenberg, K. C., D. R. Vieites, A. Van der Meijden, F. Glaw, D. C. Cannatella & M. Vences (2008): Patterns of endemism and species richness in Malagasy cophyline frogs support a key role of mountainous areas for speciation. – Evo-lution, 62: 1890–1907.

a tarzan yell for conservation: a new chameleon, calumma ... · madagascar is the ongoing rapid...

Documents